谷歌人工智能文章中英文對(duì)照

1、非專業(yè)人士翻譯，如有錯(cuò)誤請(qǐng)諒解。GooglesAIReasonsItsWayaroundtheLondonUnderground谷歌人工智能推導(dǎo)出環(huán)繞倫敦地鐵系統(tǒng)的路線DeepMind’slatesttechniqueusesexternalmemytosolvetasksthatrequirelogicreasoning—asteptowardmehumanlikeAI深度思維最新技術(shù)使用了外部存儲(chǔ)來(lái)解決需要邏輯思維和推理能力的任務(wù)B

2、yElizabethGibneyNaturemagazineonOctober142016伊麗莎白.吉布尼2016年10月14日發(fā)表于《自然》雜志Artificialintelligence(AI)systemsknownasneuralwkscanrecognizeimagestranslatelanguagesevenmastertheancientgameofGo.Buttheirlimitedabilitytorepresent

3、complexrelationshipsbetweendatavariableshaspreventedthemfromconqueringtasksthatrequirelogicreasoning.MEMYMAGIC記憶魔法Aneuralwklearnsbystrengtheningconnectionsbetweenvirtualneuronlikeunits.Withoutamemysuchawkmightneedtoseeas

4、pecificLondonUndegroundmapthoussoftimestolearnthebestwaytonavigatethetube.神經(jīng)網(wǎng)絡(luò)通過(guò)加強(qiáng)虛擬神經(jīng)元之間的聯(lián)系來(lái)學(xué)習(xí)。如果沒(méi)有存儲(chǔ)器，這樣一個(gè)網(wǎng)絡(luò)可能需要看一副特定的倫敦地鐵地圖數(shù)千次來(lái)學(xué)習(xí)最佳路線。DeepMindsnewsystem—whichtheycalladifferentiableneuralcomputer—canmakesenseofamapith

5、asneverseenbefe.Itfirsttrainsitsneuralwkonromlygeneratedmaplikestructures(whichcouldrepresentstationsconnectedbylinesotherrelationships)intheprocesslearninghowtostedeionsoftheserelationshipsinitsexternalmemyaswellasanswe

6、rquestionsaboutthem.ConfrontedwithanewmaptheDeepMindsystemcanwritethesenewrelationships—connectionsbetweenUndergroundstationsinoneexamplefromthepaper—tomemyrecallittoplanaroute.深度思維的新系統(tǒng)——他們稱它為微分神經(jīng)計(jì)算機(jī)——可以理解它從未見(jiàn)過(guò)的地圖。第一次訓(xùn)練神

7、經(jīng)網(wǎng)絡(luò)是在隨機(jī)生成的類似結(jié)構(gòu)的地圖上(被鐵路線鏈接的車站，或者其他關(guān)系)，在這個(gè)過(guò)程中學(xué)習(xí)如何將這些關(guān)系的描述存儲(chǔ)在它的外部存儲(chǔ)器并且回答問(wèn)題。面對(duì)一個(gè)新的地圖，深度思維的系統(tǒng)可以把這些新關(guān)系——按照一個(gè)圖紙上例子來(lái)連接各地鐵站之間的關(guān)系——寫到存儲(chǔ)器，并能夠回憶這些關(guān)系然后計(jì)劃路線。DeepMind’sAIsystemusedthesametechniquetotacklepuzzlesthatrequirereasoning.Aft